GLMP- realtime pedestrian path prediction using global and local movement patterns

Bera, Aniket; Kim, Sujeong; Randhavane, Tanmay; Pratapa, Srihari; Manocha, Dinesh

doi:10.1109/icra.2016.7487768

Cited by 66 publications

(66 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…First, we plan to better incorporate the non-holonomic constraints of vehicles into the computation of ORCA velocity sets, instead of using only changing responsibilities. Second, we can apply more sophisticated pedestrian motion models, such as GLMP [26], to improve the simulator. Finally, our POMDP model is actually a POMDP-lite [27], a subclass of POMDPs where the hidden state variables are constant or only change deterministically.…”

Section: Discussionmentioning

confidence: 99%

PORCA: Modeling and Planning for Autonomous Driving Among Many Pedestrians

Luo

Cai

Bera

et al. 2018

IEEE Robot. Autom. Lett.

Self Cite

157

View full text Add to dashboard Cite

This paper presents a planning system for autonomous driving among many pedestrians. A key ingredient of our approach is PORCA, a pedestrian motion prediction model that accounts for both a pedestrian's global navigation intention and local interactions with the vehicle and other pedestrians. Unfortunately, the autonomous vehicle does not know the pedestrians' intentions a priori and requires a planning algorithm that hedges against the uncertainty in pedestrian intentions. Our planning system combines a POMDP algorithm with the pedestrian motion model and runs in real time. Experiments show that it enables a robot scooter to drive safely, efficiently, and smoothly in a crowd with a density of nearly one person per square meter.

show abstract

Section: Discussionmentioning

confidence: 99%

PORCA: Modeling and Planning for Autonomous Driving Among Many Pedestrians

Luo

Cai

Bera

et al. 2018

IEEE Robot. Autom. Lett.

Self Cite

157

View full text Add to dashboard Cite

show abstract

“…Path prediction and gait analysis: There have been numerous studies on predicting the trajectories of pedestrians to prevent collisions and improve surround vehicle safety. These methods generally ignore high-level semantics (such as pedestrian intent) and predict the paths based on low level cues alone [11]- [15]. Intent analysis: The aim of such studies is to make an estimate of the pedestrians' intention in the near future, so as to take appropriate measures to reduce risk of collision.…”

Section: Related Workmentioning

confidence: 99%

When Vehicles <italic>See</italic> Pedestrians With Phones: A Multicue Framework for Recognizing Phone-Based Activities of Pedestrians

Rangesh

Trivedi

2018

IEEE Trans. Intell. Veh.

View full text Add to dashboard Cite

The intelligent vehicle community has devoted considerable efforts to model driver behavior, and in particular to detect and overcome driver distraction in an effort to reduce accidents caused by driver negligence. However, as the domain increasingly shifts towards autonomous and semi-autonomous solutions, the driver is no longer integral to the decision making process, indicating a need to refocus efforts elsewhere. To this end, we propose to study pedestrian distraction instead. In particular, we focus on detecting pedestrians who are engaged in secondary activities involving their cellphones and similar handheld multimedia devices from a purely vision-based standpoint. To achieve this objective, we propose a pipeline incorporating articulated human pose estimation, followed by a soft object label transfer from an ensemble of exemplar SVMs trained on the nearest neighbors in pose feature space. We additionally incorporate head gaze features and prior pose information to carry out cellphone related pedestrian activity recognition. Finally, we offer a method to reliably track the articulated pose of a pedestrian through a sequence of images using a particle filter with a Gaussian Process Dynamical Model (GPDM), which can then be used to estimate sequentially varying activity scores at a very low computational cost. The entire framework is fast (especially for sequential data) and accurate, and easily extensible to include other secondary activities and sources of distraction.

show abstract

“…This work assumes that there is a system other than a robot, like the one presented in [16,17], that tracks passive agents, predicts their future behaviors, and presents them in a general format, used in this paper. We are given a list of the passive agents and each passive agent is considered to be a disc or sphere having a radius and a trajectory.…”

Section: Passive Agent Representationmentioning

confidence: 99%

“…In most of the applications of navigating an agent, the moving obstacles has free will, and their future behavior is only partially predictable (if at all). When facing such situation, obstacle's path must be predicted using prediction techniques such as the ones presented in [14][15][16][17]. When the on-line path prediction is used to plan motion, it is likely that the model of the future that is obtained will have limited time duration.…”

Section: Introductionmentioning

confidence: 99%

Collision Avoidance from Multiple Passive Agents with Partially Predictable Behavior

et al. 2017

View full text Add to dashboard Cite

Navigating a robot in a dynamic environment is a challenging task, especially when the behavior of other agents such as pedestrians, is only partially predictable. Also, the kinodynamic constraints on robot motion add an extra challenge. This paper proposes a novel navigational strategy for collision avoidance of a kinodynamically constrained robot from multiple moving passive agents with partially predictable behavior. Specifically, this paper presents a new approach to identify the set of control inputs to the robot, named control obstacle, which leads it towards a collision with a passive agent moving along an arbitrary path. The proposed method is developed by generalizing the concept of nonlinear velocity obstacle (NLVO), which is used to avoid collision with a passive agent, and takes into account the kinodynamic constraints on robot motion. Further, it formulates the navigational problem as an optimization problem, which allows the robot to make a safe decision in the presence of various sources of unmodelled uncertainties. Finally, the performance of the algorithm is evaluated for different parameters and is compared to existing velocity obstacle-based approaches. The simulated experiments show the excellent performance of the proposed approach in term of computation time and success rate.

show abstract

GLMP- realtime pedestrian path prediction using global and local movement patterns

Cited by 66 publications

References 23 publications

PORCA: Modeling and Planning for Autonomous Driving Among Many Pedestrians

PORCA: Modeling and Planning for Autonomous Driving Among Many Pedestrians

When Vehicles <italic>See</italic> Pedestrians With Phones: A Multicue Framework for Recognizing Phone-Based Activities of Pedestrians

Collision Avoidance from Multiple Passive Agents with Partially Predictable Behavior

Contact Info

Product

Resources

About